The difficulty with such volumes is that the through-slice resolution is typically 5 mm, which makes the in-slice images look good (very high SNR), but makes it hard to match them to a template dataset at 1 mm resolution. It is possible to resample them to a 1 mm grid in all 3 directions ("isotropic"), and then try the standard alignment tools. However, segmentation will be inaccurate since the resolution in the z direction will still be lousy.
If I were trying this, I would assemble the slices into a 3D dataset, then try the @SSwarper command to see how well it works. Probably a fair amount of swearing at the computer will be involved along the way.